Longitudinal brain dynamics and prediction of craving changes in individuals with methamphetamine use disorder.

Individuals with methamphetamine use disorder (MUD) exhibited functional and structural brain recovery during abstinence. However, longitudinal neural changes of large-scale brain networks after prolonged abstinence remain unclear. Here, energy landscape analysis was applied to resting-state fMRI to characterize dominant activity patterns in 41 healthy controls and 40 individuals with MUD at baseline and after prolonged abstinence. We compared the dynamic characteristics of these patterns across groups and stages. Support vector regression model was constructed using baseline dynamic indices to predict craving changes after prolonged abstinence. We found that individuals with MUD exhibited rigid brain dynamics, which were correlated with abnormal brain network coordination and behaviors. Following prolonged abstinence, neural flexibility showed partial improvement, and dynamic indices could predict longitudinal craving changes. We revealed atypical brain network dynamics and their partial recovery during prolonged abstinence in MUD and suggested that dynamic indices could serve as biomarkers for predicting craving changes.